The prior art has long searched for a control system for a vehicle safety device capable of discriminating between crash events requiring actuation of the safety device and transitory mechanical inputs in response to which actuation of the safety device would be either unwarranted, undesirable or even increasingly hazardous to either the vehicle or its occupants. Current systems typically employ inertia-based accelerometers, either electromechanical "ball-in-tube" sensors or piezoelectric accelerometers operating at a frequency well below 1 Khz, to provide the system with vehicle acceleration data from which crash discrimination "measures" are derived for ultimate comparison with one or more thresholds. And, in the context of front-impact crash discrimination, such inertia-based accelerometers have benefitted from the relatively long period--perhaps 30 msec or longer--within which to provide the necessary data for crash discrimination (either through the closing of a damped acceleration-responsive "switch" or in the form of a set of digital acceleration data from inertia-based piezoelectric accelerometers). Even so, current crash discrimination algorithms used in connection with data generated with inertia-based piezoelectric accelerometers typically assume an incomplete data set with which to decide whether to actuate the safety device.
In the context of side-impact crash discrimination, the prior art is faced with a particularly limited time frame within which to decide whether to actuate the safety device--perhaps as little as 5 msec. Such a short time frame is typically inadequate to close the mechanical contacts of an electromechanical acceleration sensor. The short time frame is also generally insufficient for inertia-based piezoelectric accelerometers to provide an adequate data set for crash discrimination, thereby greatly increasing the likelihood of improper actuation of the safety device. These small and incomplete data sets may further render such piezoelectric accelerometer-based systems vulnerable to excessive. EMI and "rough road" inputs to the vehicle.
Still further, such piezoelectric accelerometer-based systems must account for the effects of temperature and component aging on accelerometer output, e.g., the accelerometer's offset and sensitivity. The resulting requirement for frequent recalibration--and the correlative requirement for the complicated methods and apparatus used for such recalibration--become increasingly important to ensure proper system response when facing side-impact crashes, with their reduced discrimination window and correspondingly greater reliance on the accuracy of the available data.
Finally, known inertia-based accelerometers typically sense only those inputs having force components applied along a sensing axis, with substantially reduced "cross-axis" sensitivity. Accordingly, great care must be exercised when installing such accelerometers within the vehicle so as to obtain the desired alignment of its sensing axis therein. Even when properly aligned, however, significant cross-axis impacts still may not produce a sufficient output generate